Banned for running OpenAI models in Claude Code? Anthropic Clarifies It Was an Error

ChrisGPT · x · 2026-08-09

Recently, a user was suspended by Anthropic after following a setup to run OpenAI models through the Claude Code harness, sparking concerns that Anthropic was penalizing users for mixing competing models.

Boris Cherny, the creator of Claude Code, stepped in to clarify that this is not Anthropic's policy. He stated that a completely separate account classifier almost certainly triggered the suspension and that the team is investigating. Tibo responded by emphasizing that 'freedom of harness is important' and users should decide which model works best. Boris jokingly tried to recruit him to Anthropic, but Tibo declined, teasing that he is 'way too excited about the next couple of weeks of releases.'

Related event: Anthropic Reverses Ban on Developers Using Claude Code with Rival Models(2 posts)→

Original post →

More from coding & agent

coding & agent channel →